随着经济的不断发展，股票市场逐渐成为金融交易领域的重要场所。而股票市场中股票价格的波动性较大，且受到非线性因素的影响明显，对于股票价格的预测一直以来都是研究者和投资者关注的重点。
       本文考察了目前主流的股票价格预测模型在A股市场上的预测性能。主要工作分为两部分。第一部分为股票多因子的特征筛选。基于股票的基本交易数据与各类特征因子，通过XGBoost模型建立股票特征因子与股票价格波动率之间的关系。利用股票特征因子在预测模型中的重要性，实现特征因子降维。第二部分，主流股票价格预测模型在A股股票上的预测性能实证。基于不同行业多支股票的日频数据，利用滑动窗口思想进行预处理，考察卷积神经网络，长短期记忆网络，门控循环单元网络在不同大小的记忆窗口下的预测性能。通过实践表明卷积神经网络模型在股价预测上预测效果优于长短期记忆网络以及门控循环单元网络，且卷积神经网络相对于记忆窗口大小不敏感，各神经网络模型在股票价格的短线预测上普遍优于长线预测。

        With the development of economy, the stock market has gradually become an important place in the field of financial transactions. However, stock price in the stock market are highly volatile and are obviously affected by nonlinear factors. The prediction of stock prices is always the focus of researchers and investors.
        This article examines the prediction performance of the current mainstream stock price prediction models in A-share market. The main work is divided into two parts. The first part is the feature screening of stock multi-factors. Based on the basic trading data of stocks and various characteristic factors, the relationship between stock characteristic factors and stock price volatility is established through the XGBoost model. Using the importance of stock characteristic factors in the prediction model, the dimension reduction of characteristic factors is realized.  The second part is the examination of the mainstream stock price prediction models’ prediction performance on A-share. Based on the daily frequency data of several stocks in different industries, the sliding window idea is used for preprocessing, and the prediction performance of convolutional neural network, long short-term memory network and gated recurrent unit network under different memory windows is investigated. Practice shows that the convolutional neural network model is better than the long-term and short-term memory network and the gated recurrent unit network in stock price prediction, and the convolutional neural network is insensitive to the size of the memory window. Each model is generally better than the long-term prediction in the short-term prediction of stock prices.